Research On Bearing Fault Diagnosis Of Printing Machine Based On Auxiliary Generating Deep Belief Network

Bearing running performance is essential to the safe and stable operation of rotating machinery.The printing machine is one of the typical rotating machinery,in which the running performance of the main components of the printing machine bearing will indirectly affect the printing quality.In order to solve the problems of the traditional fault diagnosis methods,such as low efficiency with large samples,and the diagnostic accuracy is easily affected by the quality of the collected signals.In this paper,a method of printing press bearing fault diagnosis based on quantum genetic optimization and auxiliary generation of deep belief network is proposed.This method mainly balances the original data set by the auxiliary generation algorithm,and uses the deep belief network to extract the hidden fault features from the samples and identify the fault types.Then,the model weights are optimized adaptively by using quantum genetic algorithm.The specific research contents are as follows:Considering the complexity of the traditional fault diagnosis methods for fault signal feature extraction,and limitation of the signal processing technology and expert experience,this paper studies the method of feature extraction and pattern recognition for bearing faults using deep belief network.The self-learning ability of the deep belief network is used to learn the fault features in the bearing vibration signal,and then use the extracted fault features to identify the type of fault bearing.The bearing data set was used to verify the versatility and effectiveness of the method,with an average diagnostic accuracy of 91.83%.The results show that the method has higher diagnostic accuracy and weakens the need for manual extraction of fault features.Considering the problem of low diagnostic accuracy of traditional deep belief network bearing fault diagnosis methods when dealing with unbalanced data sets,this paper studies a fault diagnosis method based on auxiliary generation of deep belief network.This method generates some samples according to the original sample distribution through the auxiliary generation network,and then amplifies the unbalanced data set to improve training efficiency and diagnostic accuracy.The method is verified in the constructed bearing unbalanced data set The experimental results show that the method has achieved a good diagnostic accuracy compared with the traditional method,and has certain research valueConsidering the high requirements on parameter selection of traditional fault diagnosis methods based on deep learning algorithms and manual selection of model parameters,the quantum genetic algorithm is introduced,and a method of printing press bearing fault diagnosis based on quantum genetic optimization and auxiliary generation of deep belief network is proposed is proposed.This method can adaptively select model parameters,and accurately and efficiently search for the optimal solution of the parameters,avoiding the phenomenon of accuracy degradation caused by the parameter selection.This method is applied to the fault diagnosis of printing machine bearings,and the superiority of this method is verified through field experiments.The accuracy of fault diagnosis for unbalanced data sets can reach 84.99%.A software for bearing fault diagnosis system of printing press is designed and completed according to the research model of this paper.